Effective coordination requires that the motor system predict proper movements. To make these predictions, the cerebellum integrates sensorimotor information and motor errors and, through a process of error-driven learning, builds up feed-forward models of movement. Decades of cerebellar research have clarified a highly stereotyped circuit, identified roles for particular circuit elements, and suggested cellular mechanisms that might account for associative learning. However, fundamental questions remain unanswered. Where within the cerebellar circuit do changes occur during cerebellum-dependent forms of motor learning? How, at a cellular and circuit level, do motor errors drive these changes? And finally, how do circuit changes alter cerebellum-dependent behavior? The following specific aims will be addressed in the project. In Specific Aim 1, we will test the hypothesis that pauses in PN firing, evoked directly or indirectly by optogenetic stimuli, trigger rapid, highly stereotyped movements. Using high speed videography and motion tracking we will measure movement trajectories and speeds in response to activation or inhibition in various cerebellar neurons. We will also make electrophysiological recordings from cerebellar neurons in awake mice to examine the effects of manipulating PN excitability on the circuit. In Specific Aim 2 we will test whether associative motor learning can be driven by pairing sensory stimuli with optogenetically- elicited reductions or increases in PN firing. In Specific Aim 3 we will use n vivo electrophysiology to explore neural mechanisms of learning at key sites in the cerebellar circuit and to determine how error signals contribute to this learning. These interrelated aims make use of a novel behavioral preparation applying optogenetic, electrophysiological, and behavioral methods to awake mice in order to answer fundamental questions about cerebellar physiology. Together, the proposed experiments are designed to resolve issues that have been debated for decades within the cerebellar field. We expect that our results will yield a much improved understanding of basic cerebellar physiology and resolve some long-standing mysteries regarding cerebellum-dependent learning. In addition, these findings are likely to provide conceptual insights into cerebellar dysfunction caused by inherited and sporadic forms of ataxia and dystonia.